个人所得税的公平性体现

40THE CHINESE ECONOMY

The Chinese Economy, vol. 39, no. 2, March–April 2006, pp. 40–56.

© 2006 M.E. Sharpe, Inc. All rights reserved.

ISSN 1097–1475 / 2006 $9.50 + 0.00.

HOWELL H. ZEE AND FARHAN HAMEED

Reforming China’s Personal

Income Tax

Abstract: This article derives the distribution of wages in China on the basisof the Pareto distribution, and employs it to simulate the revenue and dis-tributive consequences of various options of restructuring the personal in-come tax schedule as applied to wage income. The obtained simulation resultsare interesting and plausible and could thus inform policy deliberations onreforming the personal income tax in China.Reforming China’s personal income tax (PIT), which was introduced in itspresent form in 1994, has been a focus of active research and policy debateswithin China in recent years. In a press conference held by the InformationOffice of the State Council in January 2004, Xie Xuren, commissioner of theState Administration for Taxation, indicated that PIT reform would be part ofa comprehensive tax reform program being launched in 2004 that would alsoinclude reforms of the value-added tax and corporate income tax.1While all aspects of the PIT are possible targets for reform, the focus ofthis article is on its rate schedule as applied to wages and salaries. Since nodisaggregated data on PIT revenue collection by rate band are yet availablein China, the article shows how the Pareto distribution could be used to assistpolicy deliberations on the restructuring of such a schedule. The purpose of

the article is not to make specific reform recommendations, but to illustrate

Howell H. Zee is at the Fiscal Affairs Department, International Monetary Fund(IMF), Washington, DC. Farhan Hameed is also affiliated with the IMF.

The views expressed herein are those of the authors; they do not necessarily reflectIMF policy and should not be reported as representing the views of the IMF. Theauthors thank Ray Brooks and Reed Shuldiner for their helpful comments. The usualdisclaimer applies.

40

MARCH–APRIL 200641

the possible value of applying a particular but plausible analytical model tohelp address a problem that would otherwise be difficult to analyze in viewof the data limitations. A brief history and the present structure of China’sPIT, together with its revenue collection in 2002, are first provided in thenext section, followed by a description of the analytical model in the thirdsection. The fourth section illustrates the way the estimated parameter valuesfrom the analytical model could be used to experiment with alternative rateschedules, and also discusses their economic consequences. Some brief re-marks in the final section conclude the article.

History, Present Structure, and Revenue Collection of

PIT History

History

The present PIT in China has its roots in three different taxes on personalincome that were introduced in the 1980s:2

The PIT, which became effective in 1980 (henceforth the old PIT), wasimposed on specific categories of worldwide income of residents and in-come of nonresidents sourced in China. Wages and salaries were subject totax on a six-band progressive marginal rate schedule ranging from 5 percentto 45 percent, with a basic monthly allowance of RMB800.3 All other cat-egories of taxable income, namely, (1) income from personal and profes-sional services, (2) royalties, (3) interest (except on saving deposits in statebanks and credit cooperatives), dividends, and other distributed profits, and

(4) rental income, were taxed at a flat 20 percent rate.

The individual commercial and production enterprise (ICPE) incometax,4 which became effective in 1986, was imposed on the business incomeof ICPEs according to a ten-band progressive marginal rate schedule rangingfrom 7 percent to 60 percent. In addition, a surtax of 10 percent to 40 per-cent, determined by the local authorities, was imposed on the tax liabilityderived from taxable income in excess of RMB50,000.

The personal income regulatory tax, which became effective in 1987,was imposed on the following specific categories of worldwide income ofresidents (who would no longer be subject to the old PIT):5 (1) wages andsalaries; (2) income from contracted or leased operations; (3) income frompersonal and professional services; (4) remunerations of authors; (5) royal-ties; (6) interest (except on saving deposits in state banks and credit coopera-tives), dividends, and other distributed profits; and (7) rental income. Theaggregate income from categories (1) through (3) and (7) was subject to a

five-band progressive marginal rate schedule ranging from 20 percent to 60

42THE CHINESE ECONOMY

percent, with a basic monthly allowance of four times the stipulated basic(average) monthly wage (RMB100–RMB115) that varied across four speci-fied regions. Income from categories (4) through (6) was taxed at a flat 20percent rate.

Present Structure

The present PIT law, which was promulgated in 1993 (and became effectivein 1994) as an amendment to the old PIT law, is in effect an amalgamation ofthe above three taxes, covering both residents and nonresidents. At the sametime, further adjustments were made to the tax treatments of, and applicabletax rates on, the different categories of taxable income, which no doubt con-tributed to the complexity of the present PIT, whose structure can be brieflysummarized as follows:6

There are eleven categories of taxable income: (1) wages and salaries(taxable monthly); (2) business income of ICPEs (taxable yearly); (3) in-come from contracted or leased operations (taxable yearly); (4) income frompersonal or professional services (taxable upon receipt if occasional, butmonthly if continuing); (5) remuneration of authors (taxable per occurrenceof publication); (6) royalties (taxable per issuance of licensing rights); (7)interest, dividends, and other distributed profits (taxable upon receipt);7 (8)rental income (taxable monthly); (9) capital gains from real and financialassets (taxable per realization); (10) incidental income (taxable upon receipt);and (11) other categories as specified by the Ministry of Finance.

Allowances and deductions for each category of taxable income are speci-fied separately. Most notably, a basic monthly allowance of RMB800 is pro-vided to domestic residents for category (1) income.8 There are no otherpersonal allowances available under this income category.9

Category (1) income is subject to tax according to a nine-rate progres-sive schedule, as shown in Table 1.

Separate progressive schedules are applied to income categories (2), (3),and (4). Category (5) income is taxed at a flat rate of 14 percent, while in-come in all other categories is taxed at 20 percent (except rental income frompersonal residential properties, which is taxed at 10 percent).

Revenue Collection

While on a rising trend in recent years, revenue from the PIT is still fairlyinconsequential in China at present. As shown in Table 2, PIT revenueamounted to only about 1.2 percent of gross domestic product (GDP) (less

than 7 percent of total tax revenue) in 2002. The bulk of this revenue was

MARCH–APRIL 200643

Table 1Nine-Rate Progressive Schedule of Category (1) Income

Monthly taxable income (net of basic monthly allowance)Marginal tax rate (%)Not exceeding RMB5005Above RMB500 but not exceeding RMB2,00010Above RMB2,000 but not exceeding RMB5,00015Above RMB5,000 but not exceeding RMB20,00020Above RMB20,000 but not exceeding RMB40,00025Above RMB40,000 but not exceeding RMB60,00030Above RMB60,000 but not exceeding RMB80,00035Above RMB80,000 but not exceeding RMB100,00040Above RMB100,00045

derived from three categories of income: wages and salaries (46 percent oftotal), interest and dividends (32 percent), and business income of the ICPEs(15 percent).

The bulk of the PIT revenue accrued to the provincial governments as partof the comprehensive tax reform that took place in 1994,10 an arrangementthat lasted until 1999, when it was stipulated that the PIT revenue derivedfrom deposit interest would accrue to the central government. A change tothis revenue sharing occurred in 2002, when it was decided that the excess ofPIT revenue (inclusive of the revenue from deposit interest) above the 2001level was to be shared between the central and provincial governments—inthe proportions of 50 percent and 50 percent in 2002, and 60 percent and 40percent in 2003.

Application of the Pareto Distribution

A central theme in reforming China’s PIT will undoubtedly be simplificationwith respect to both its base and rate schedules. The base issue will not bepursued further in this article.11 With regard to the rate schedules, clearly themost important schedule on which to focus one’s attention is the one that isapplied to wages and salaries. To analyze the revenue and distributive conse-quences of restructuring this schedule, detailed data on the PIT revenue col-lected by rate band are needed. If the restructuring is free to determine newband widths that bear no relationship to existing ones, data on national wagedistribution would also be necessary. Neither type of data at a sufficiently

disaggregated level is currently available in China. The extent of available

44THE CHINESE ECONOMY

Table 2

Composition of PIT Revenue, 2002

Amount inPercent

RMB bn of total

Wages and salaries56.146.4Of which: taxable at ≥ 25 percent9.78.0Business income of ICPEs18.515.3Interest, dividends, and bonuses38.431.7Of which: deposit interest30.124.8Other income8.16.7Total121.1100.0

Memorandum item

Total PIT revenue

As percent of total tax revenue6.9

As percent of GDP1.2

Source: State Administration of Taxation.

disaggregation of PIT revenue data is shown in Table 2: only revenue col-lected from the 25 percent rate band and above can be separately identifiedfrom the total in 2002 (this is the only year for which such a breakdown isavailable).

Lacking the requisite distribution data, the underlying wage distributionin China could be constructed on the assumption that it follows the Paretodistribution, which has been shown to be a reasonable assumption in manyempirical studies on income distribution in a large number of developed anddeveloping countries.12 The constructed distribution could then be used tosimulate the collection of PIT revenue by rate band of the progressive PITschedule applied to wages and salaries. The parameters of the model areestimated in such a way that the constructed distribution produces an out-come that is consistent with both the aggregate PIT revenue collection andthe subtotal amount collected at 25 percent and above in 2002.

Conceptual Framework

With the Pareto distribution, the distribution of the population over the rangeof wage levels (w), denoted by f(w), that is equal or above

some positive

MARCH–APRIL 200645

threshold w0, expressed as a fraction of total population at each w, is assumedto take the following functional form:

f(w) = β w-α-1, w ≥ w0 > 0,(1)

where β º α wa

coefficient. Hence, this distribution has two parameters: 0 > 0 is a constant and α > 1 is commonly known as the Paretoα and w

equation (1), the cumulative distribution function, denoted by F(w0. Given), of thePareto distribution, which gives the proportion of the population with wagesno greater than w, takes on a particularly simple form. By integrating f(w)over the wage interval [w0, w], one obtains

F(w) = 1 – (w0/w)α.(2)

Hence, the proportion of the population with wages above w is simply 1 –F(w) = (wα0/w). It follows that, if the size of the population is given by N,the number of wage earners with wages above w, denoted by n(w), wouldthen be

n(w) = N (w0/w)α,(3)

which is the most commonly cited formula associated with the Paretodistribution.

The average wage income for the entire population under the Pareto dis-tribution turns out to be13

v = w0 α/(α – 1),(4)

which says that it is proportional to the threshold income w

relatively simple matter with this distribution to ascertain both the number of0. It is also a

wage earners and the total wage income over any wage interval [a, b], whereb > a ≥ w0. The number of wage earners with wage income in this interval,denoted by na,b, is simply the product of N and the definite integral of f(w) over the interval, that is,

na,b = N β (a-α – b-α)/α.(5)

In the similar vein, total wage income in this interval, denoted by wbe the product of N and the definite integral of [w a,b, would

f(w)] over the interval,

that is,

wa,b = N β (a1-α – b1-α)/(α – 1).(6)

Hence, the average wage income in this interval is

46THE CHINESE ECONOMY

va,b ≡ wa,b/na,b = [(a1-α – b1-α)/(a-α – b-α)] [α/(α – 1)].(7)

Model Implementation

The objective is to ascertain the PIT revenue from wage income in each ofthe nine rate bands, for which actual data are not available. Since the basicmonthly allowance of RMB800 is equivalent to having a zero-rate band forincome below the said allowance, one could regard the rate schedule as hav-ing ten rate bands, starting with the zero percent rate and ending with the topmarginal rate of 45 percent. This schedule is given in column 1 of Table 3,with column 2 indicating the width of each band expressed in terms of an-nual wage income.

There are initially three unknown parameters: α, N, and w

w0. However, thegiven monthly allowance permits one to set 0 = 9,600, leaving only twoparameters to be estimated. Figure 1 shows how total PIT revenue behaves in

response to different combinations of α and N. The solution values of thesetwo parameters are found through numerical simulations so that the simu-lated PIT revenue outcome corresponds to the actual PIT revenue collectionin 2002 (see Table 2), for both (1) the global amount of RMB56.1 billion and

(2) the subtotal amount of RMB9.7 billion, which represents the PIT revenuecollected from rate bands from 25 percent to 45 percent. In Figure 2, thecurve G is an iso-PIT revenue contour that traces out the various combina-tions of α and N that would achieve the actual global PIT revenue target,while the curve M is another iso-PIT revenue contour that traces out variouscombinations of α and N that would achieve the actual intermediate PITrevenue target. The point where the two curves intersect provides the (lo-cally) unique combination of α and N that would achieve the global andintermediate targets simultaneously. Specifically, the simulation producedthe following solution values: α = 1.882197575 and N = 48,613,963.

Using these solution values of a and N, columns 3 and 5 of Table 3 can bederived by applying equations (5) and (7), respectively. Column 6 then com-putes the PIT on the average wage in each rate band by applying the rateschedule given in column 1. The total PIT revenue in a given rate band issimply the product of the number of taxpayers and the PIT in that rate band,shown in column 7. Column 9 shows the average PIT rate in each rate band.

The simulation results imply that, in 2002, China had about 49 millionPIT payers14 with an average annual wage income of about RMB20,000. Theaverage PIT rate progresses from about 1 percent (in the 5 percent rate band)to about 31 percent (in the 45 percent top rate band), yielding an overallaverage PIT of about 5.7 percent.15

Not surprisingly, while about 91 percent

3 elb

Tae7geat)rra% (x Aveat)latot foT6 IP%( latTo ) B M R ( -rve5e)ag aBnowM RTeg(IPaeg4)BraegMAve waR( )latot fo r3e%y(apx)Tar e b m u N(l)leba2eBagxnuMnRTaawa ( la1netigarra x)

Ma%

t(015236417.........—148492580112223048677251.........—633983211222—052396410086580109928990054,,,,,,,,,749764620298847490156948611,,,,,,,,,714330057645268343313680286,,,,,,,3336421111—622724797189418917188616741,,,,,,,35421251631001234—335769683124347099945308389,,,,,,,,,115026369124138371135803,,1196122.....—9072053—————554893052089158197305486875,,,,,,,,,0412555212936711840,,,,943121008000000000000000006666666662,,,,,,,,,5953999999,–1364826010–––247925111––––,400011111,6660000–1,,,95366661–,,,,01139999606482,62479,69902,1050505050511223344MARCH–APRIL 200647)de7nu.i5tnoc(09102..0701109100000,00,6066,24,4715,66,59932,0209102..000109122012,70,1160,184latTo

Simulated Distributio

n of

PIT

Rev

enue from Wage Income, 2002

48THE CHINESE ECONOMY

)

deunitnoc

( 3 elb

Tae7geat)rra%

(x Aveat)latot foT6 IP%( latTo ) B M R ( -rve5e)ag aBnowM RTeg(IPaeg4e)gBraMAve waR( )latot fo r3e%y(apx)Tar e b m u N(l2)lebaeBagxnuMnRTaawa ( la1netigarra x)

Ma%

t(1211000000,,006662,,474516,,695.369,316,84 =N ;575791288.1 = α :seulav reetuenemvarear pT Ide PtlaaumittcsAE.008BRM f.od encanb aewtaorl lea myalhs tenhot mf oe thit mgiln irtenewso.)el 2re pheletr b sdeetmTa rime.t) osn12r(.fi 1( dn tenmav aut.toble osnboc ea.r )yc en6lrdlenzia (ipaen c hen5t.itvm)f4ne ih.gu5i tltd (rocrie atwzson e tiac l ,edmsinp e5ealudc2mulr nunaonefond e)capoca b / n h3( 5s,c )eit ssn6a2d ned (ragnem .taaaunt.mwbrl .. omnso)bedr1 e ))lcucnf etne57a lraslaoga((nfoebdbr iocp nenng t elwf5amtu oorceiiao4ab Tao dlttu rebaamdo e emrfto ihauu othTardcxqqeratIrfet (rsaeehtpfP t e gggeeolT ceelIltfnnhh taPulaoniiitti no r s mrre hsieevc Ttuusussaa lvvee Id ddddrooruPideeeee wetttttl pddTnealuuuuupeeIuapppppummPv etu ctmmmmmemmlra AcooooohutuTI CCSSP A CC C T To ............123456789012

111Memorandum items:

Notes:

s

egaW morf eun

eveR TI

Plat

Todetalum

i

S

1

e

rugiFMARCH–APRIL 200649)snoillim ni0( 4

.9N009.1α9988)nauy fo snoillib ni( segaw morf eunever TIP latoT

50THE CHINESE ECONOMY

sruotnoC eun

eveR TI

-PosI

2

e

r

u

g

i

F00.2081.

061.04.100000054321)

snoillim ni( N60α

MARCH–APRIL 200651

of the taxpayers are located in the 5 percent and 10 percent rate bands, thesetaxpayers only contributed about just below 30 percent of total PIT revenue.Restructuring the PIT Rate Schedule on Wage Income

The estimated model above can be used to simulate the revenue impact of allpossible changes to the rate structure, including setting completely new widthsfor any rate band one may wish to consider. However, for illustrative pur-poses and ease of comparing the distribution of tax burdens across rate bandsunder alternative rate structures with that under the existing rate structure,the existing band widths for all rates have been retained for the simulationexercise illustrated below. In other words, a reduction in the number of ratesis achieved by simply collapsing existing rates without defining new widthsfor the rate bands.

Table 4 presents three broadly revenue-neutral rate-restructuring options:

(1) a three-rate option comprising 10 percent, 20 percent, and 30 percent; (2)a two-rate option comprising 15 percent and 25 percent; and (3) a flat-rateoption. Under all options, it is assumed that the basic allowance has beenraised to RMB1,300 per month, or RMB15,600 per year. This raise servestwo useful purposes: first, it removes all those in the existing 5 percent rateband—comprising almost 60 percent of the total number of current PIT pay-ers—from the tax net, thus benefiting the least well off as well as reducingcompliance and collection costs; and second, it allows the marginal tax rateon those in the existing 10 percent rate band to be raised (if necessary) with-out necessarily increasing their average tax rate. Coincidentally, a basic al-lowance of RMB1,300 per month also happens to broadly maintain the realvalue of the basic allowance that was provided in 1987 under the personalincome regulatory tax (see above). Raising the basic allowance further, say,to eliminate the existing 10 percent rate band, would be very costly in rev-enue terms (losing close to 30 percent of the present total PIT collection) andwould require a significant jump in the marginal tax rates on all other tax-payers for a revenue-neutral outcome. Under the three-rate option, taxpayersin the existing 5 percent and 10 percent rate bands would be made better off,while those in the existing 15 percent rate band would see their average taxrate increase only slightly (by 0.7 percentage points). The taxpayers whowould be most adversely affected under this option would be those in theexisting 20 percent and 25 percent rate bands, but even for these taxpayersaverage tax rates would rise by at most seven percentage points. Under thetwo-rate option, no taxpayer would see his average tax rate rise by more thanthree percentage points, and those in the existing 10 percent rate band would

do no worse than before. In sum, either option would leave taxpayers in the

52THE CHINESE ECONOMY

t

4 elb

TaseerggutencgetiWatsuaa)rrrit nxse vx%(oE eAatt eraludeehgetcaa)rre Seevx%(ttAataraR t) TaenlIFTub : In P3PeBe vhneMtorR(i tropfOlae ntsinga)rr oax%(itMatpO gneigetruaa)rrte %(csvxueAatrttasr) eenRTubTwoIn eBe: Pvv2eMi rtanR(orilttpaesntuOiga)rlr lx%(Ia aeMtvitanreegettaa)lrrAse % fevx(otAat sra )teencearTub phInTPeBv: eMI

m1rR( nneolid

taerpnt)uOigarBr %ax(

xMatTa

dn1ae egaunew) llKeba vauBexnM

RnTaaR

(—015236417.........148492580112223——62625788........414666661111111——751...158536......4771114100065——1717171717171717——19624913........49623341222422——044...622...750...3591115210085——1515252525252525——722406......91..29968812228922——745764......967...842621007125——001020330303303025.6.1566669...466.9990,6..98262,1.394791–53624––––6.61–––66669l....9–666....99990a–6t.53948262,0913624791Toa tes yllaicifitra si dnab etar pot eht fo timil reppu ehT2 .dnab etar-orez a sa 008BMR fo ecnawolla ylhtnom cisab e.thit msiel traerowporl osctin Is1e :smeitt

o2

N.

1

MARCH–APRIL 200653

existing lowest two rate bands at least as well off, with most of them actuallymuch better off, than under the existing rate schedule. Taxpayers in the highrate bands under the existing rate structure would obviously be made betteroff by either reform option under discussion (more so by the two-rate thanthe three-rate option).

With the basic monthly allowance set at RMB1,300, the flat tax rate thatwould produce a broadly revenue-neutral outcome would be 17 percent.16However, in this scenario, taxpayers in the existing 10 percent rate bandwould actually see their average tax rate rise by 0.5 percentage point; thosein the existing 15 percent rate band would also be more adversely affectedthan by either of the two options discussed above. Hence, the burden of theflat tax, as expected, falls largely on the lower-income taxpayers. Figure 3provides a comparative look at the different profiles of average tax ratesunder the various scenarios.

The three restructuring options illustrated above are chosen because theyseem to generate theoretically interesting outcomes. As noted earlier, thereare countless other restructuring options one could investigate using the simu-lation model by varying the level of the basic monthly allowance, the num-ber of tax rates, and/or the width of each rate band.

Concluding Remarks

A crucial part of any PIT reform usually has to do with restructuring thePIT’s progressive rate schedule: perhaps to simplify it (e.g., reducing thenumber of rates), to reduce marginal rates (e.g., aligning the top marginalrate with the corporate income tax rate), and/or to improve equity (e.g., in-creasing average progressivity by raising the basic allowance). However, anyrate restructuring inevitably entails important revenue and distributive con-sequences. A proper analysis of such consequences would require detaileddistribution data on taxable income and PIT revenue collection by rate band—the sort of data that are often not available in developing countries (such asChina), thus possibly hampering the quality of reform deliberations bypolicymakers in those countries.

This article shows that the data limitation could sometimes be overcome,at least partially, by an analytical modeling of the underlying distribution ofincome, for example, by assuming that the distribution follows the well-knownPareto distribution. It is further demonstrated that the Pareto distribution canbe readily applied to the China case in a straightforward manner. The ob-tained results of this particular application seem plausible and could thusinform policy deliberations on restructuring China’s PIT rate schedule on

wage income. However, the objective of the article is not to make specific

54THE CHINESE ECONOMY

segaW no eludehc

SetaR TI

Peht rof snoitp

O gni

rutcurtseR evitartsullI evitan

retl

ArednU selifor

PetaR x

Taega

revA

3

e

r

u

g

i

F5440

3503

5d2nat) benteacrr etnp eni0rr(2uC510150.0.0.0.0.0.0.0050505032211)

tnecrep ni( etar xat egarevA35.0

MARCH–APRIL 200655

recommendations for such a restructuring, but rather to provide a relativelysimple illustration of how the Pareto distribution could be put to good ana-lytical use to address a PIT reform issue in a real country setting.

Postscript

After a protracted period of debate, China has doubled the level of the PIT’sbasic monthly allowance for wage income to RMB1,600, effective January1, 2006.

Notes

1. Xie’s statement is available at /ce/doc/cep2/200403173397.htm.

2. For an informative description of these taxes, see Li 1991.

3. Between August 1987 and the introduction of the present PIT in 1994 (effec-tively replacing the old PIT—see further below), the tax liability on wages and sala-ries of nonresidents was reduced by half.

4. An ICPE is essentially a household enterprise owned and managed by one ormore members of a household.

5. The main motivation behind the introduction of this tax—hence effecting theseparate taxation of residents and nonresidents—was the basic monthly allowanceprovided under the old PIT, which was judged to be too high for residents.

6. See the Personal Income Tax Law of the People’s Republic of China (1980, asamended); Regulations for the Implementation of the Personal Income Tax Law ofthe People’s Republic of China (1994); and various policy and administrative decreesissued by the Ministry of Finance, by the State Administration of Taxation, and byboth jointly.

7. The 1993 amendment also included an exemption of all interest from depositsand government bonds from tax. The deposit interest exemption was subsequentlyremoved in a 1999 amendment.

8. Taxpayers who do not have a domicile (habitual abode) in China but receivedomestic-source wages and salaries, or those who are domiciled in China but receiveforeign-source wages and salaries, are given an additional monthly allowance ofRMB3,200.

9. Recently, however, the local authorities of a number of “high-cost” cities, suchas Beijing, Guangzhou, and Shanghai, have also sanctioned supplemental monthlyallowances of various amounts (at a cost to their own budgets) for domestic residents,Some of the supplemental allowances even exceed the RMB800 basic monthly allow-ance as stipulated in the PIT law and regulations.

10. The main exception was the PIT collected from the oil sector, which accrued(and still accrues) to the central government.

11. For a review of relevant reform issues relating to the base of a PIT, see Zee2005.

12. For good discussions of the Pareto distribution, theory and applied, see Cowell1995 and Champernowne and Cowell 1998.

13. Note that the mean of this distribution is not defined for 1 ≥ α

.

56THE CHINESE ECONOMY

14. This would represent about 13 percent of the employed labor force (excludingthe agricultural sector), which is not an unreasonable figure, given that the basic monthlyallowance is not much below the national average monthly wage of RMB1,035 in2002. Published data on urban household income survey in the same year (see Na-tional Bureau of Statistics 2003) likewise suggest that the PIT rate schedule on wageincome covered at most only the top quintile of such households.

15. This compares with the national average of 3.5 percent (the ratio of total PITrevenue collected from wages and salaries to national total wage income). This dis-crepancy could be explained by the fact that, in the simulation model, wages earnedby those in the zero-rate band are ignored.

16. The revenue-neutral flat rate would be about 11 percent if the basic monthlyallowance is kept at RMB800.

References

Champernowne, D.G., and Frank A. Cowell. 1998. Economic Inequality and

Income Distribution. Cambridge: Cambridge University Press.

Cowell, Frank A. 1995. Measuring Inequality. Hemel Hempstead, UK: Prentice

Hall/Harvester Wheatsheaf.

Li, Jinyan. 1991. Taxation in the People’s Republic of China. New York: Praeger.National Bureau of Statistics. 2003. China Statistical Yearbook. Beijing: China

Statistics Press.

Zee, Howell H. 2005. “Personal Income Tax Reform: Concepts, Issues, and

Comparative Country Practices.” IMF working paper WP/05/87. Washington,DC: International Monetary Fund.

To order reprints, call 1-800-352-2210; outside the United States, call 717-632-3535.

本文来源：https://www.bwwdw.com/article/ds2m.html